Opposite Effects of Mn²⁺ and Zn²⁺ on PsaR-Mediated Expression of the Virulence Genes pcpA, prtA, and psaBCA of Streptococcus pneumoniae

Abstract

Homeostasis of Zn²⁺ and Mn²⁺ is important for the physiology and virulence of the human pathogen Streptococcus pneumoniae. Here, transcriptome analysis was used to determine the response of S. pneumoniae D39 to a high concentration of Zn²⁺. Interestingly, virulence genes encoding the choline binding protein PcpA, the extracellular serine protease PrtA, and the Mn²⁺ uptake system PsaBC(A) were strongly upregulated in the presence of Zn²⁺. Using random mutagenesis, a previously described Mn²⁺-responsive transcriptional repressor, PsaR, was found to mediate the observed Zn²⁺-dependent derepression. In addition, PsaR is also responsible for the Mn²⁺-dependent repression of these genes. Subsequently, we investigated how these opposite effects are mediated by the same regulator. In vitro binding of purified PsaR to the prtA, pcpA, and psaB promoters was stimulated by Mn²⁺, whereas Zn²⁺ destroyed the interaction of PsaR with its target promoters. Mutational analysis of the pcpA promoter demonstrated the presence of a PsaR operator that mediates the transcriptional effects. In conclusion, PsaR is responsible for the counteracting effects of Mn²⁺ and Zn²⁺ on the expression of several virulence genes in S. pneumoniae, suggesting that the ratio of these metal ions exerts an important influence on pneumococcal pathogenesis.

The human pathogen Streptococcus pneumoniae is a commensal of the nasopharynx, but it can become virulent and infect the lungs, middle ear, brain, and bloodstream, causing severe diseases such as pneumonia, otitis media, meningitis, or sepsis (see references 10 and 61 for reviews). Little is known about the specific environmental changes that S. pneumoniae encounters during infection of the human body and the way in which these affect virulence. One important environmental factor is probably the concentration of metal ions, which form a class of nutrients that are important for bacteria in small amounts but are often toxic in larger amounts (9, 20, 57, 58).

An important metal ion that S. pneumoniae could face is Zn²⁺, which is present in the human body in concentrations ranging from a few μM to over 100 μM (81). In the host, Zn²⁺ is of great importance for immunity, as it is necessary for proper functioning of immune cells (28), and mild Zn²⁺ deficiency severely affects immune function (72). Zn²⁺ levels are elevated during inflammation (53), and Zn²⁺ administration reduces airway infections in children in developing countries (71). Moreover, Zn²⁺ deficiency results in an increased risk of pneumococcal infection and death in mice (75) and in a lower immune response to the pneumococcal antigen PspA (76). Interestingly, several studies suggest that the use of and response to metal ions, such as Zn²⁺, Mn²⁺, and Fe²⁺, is important for the virulence and physiology of pathogenic streptococci (3, 11, 12, 18, 30, 32, 39, 40, 46, 50, 54, 55, 63, 64, 66, 67).

In S. pneumoniae several systems dedicated to the acquisition of specific metal ions have been studied. These are the PsaBCA Mn²⁺ transporter (18, 44, 50), the AdcCBA Zn²⁺ uptake system (18, 19), and three iron uptake loci (piaABCD, piuBCDA, and pit) (11, 12). The PsaBCA complex is involved in virulence, oxidative stress, penicillin stress, competence, and adhesion via interaction with human E-cadherin (2, 6, 18, 32, 48, 50, 59, 60, 79). Expression of the psaBCA genes is repressed by the DtxR family regulator PsaR in response to high Mn²⁺ concentrations (31). S. pneumoniae also contains systems involved in cation efflux (58), such as czcD, which has recently been shown to be responsible for Zn²⁺ resistance (39). The presence in S. pneumoniae of this Zn²⁺ efflux system together with the AdcCBA Zn²⁺ uptake system indicates that this pathogen has to deal with fluctuating Zn²⁺ levels in the human body.

Therefore, we explored the response of S. pneumoniae to a high Zn²⁺concentration by means of transcriptome analysis. Interestingly, Zn²⁺ was found to increase expression of several genes involved in virulence, namely, prtA, pcpA, and psaBC, which could be counteracted by Mn²⁺. Subsequent research showed that the transcriptional regulator PsaR is directly responsible for the Mn²⁺-dependent repression and the Zn²⁺-dependent derepression of these genes.

MATERIALS AND METHODS

DNA techniques, β-galactosidase assays, bacterial strains, and growth conditions.

All DNA manipulation techniques, growth conditions, and media were the same as described previously (37, 38) unless indicated otherwise. β-Galactosidase assays were performed as described previously (38).

Chelex-treated M17 was prepared by autoclaving 2% Chelex 100 resin (Bio-Rad) with M17, followed by 2 h of stirring. After removal of the resin, 50 μM CaCl₂, 50 μM MgCl₂, 5 μM FeCl₃, and 0.25% glucose were added, and the resulting medium, designated GM17chel, was used as growth medium as specified in Results. Metal ions were added as the salts ZnSO₄, MnCl₂, MgCl₂, CaCl₂, NiCl₂, CuSO₄, and FeSO₄. The strains and plasmids used or constructed in this study are listed in Table 1. Primers are listed in Table 2. Since most work presented in this paper was carried out before publication of the D39 genome sequence (43), primer sequences are based on the R6 genome sequence (27).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Reference or source
S. pneumoniae
D39	Serotype 2 strain, cps2	4, 43; laboratory of P. Hermans
R6	D39 (cps2 2538-9862) with increased transformation efficiency	27
D39nisRK	D39 ΔbgaA::nisRK; Trmpr	37
MP102	D39 ΔczcD	This work
RW100	D39 ΔpsaR	This work
RW101	D39nisRK ΔpsaR	This work
RW102	D39 ΔbgaA::PpcpA-lacZ; Tetr	This work
RW103	D39 ΔbgaA::PnrdD-lacZ; Tetr	This work
RW104	D39nisRK prtA-lacZ; Ermr	This work
RW105	D39 ΔbgaA::Pspr0276-lacZ; Tetr	This work
RW106	D39 ΔbgaA::PpsaR-lacZ; Tetr	This work
RW107	RW100 ΔbgaA::PpcpA-lacZ; Tetr	This work
RW108	RW100 ΔbgaA::PnrdD-lacZ; Tetr	This work
RW109	RW101 prtA-lacZ; Ermr	This work
RW110	RW100 ΔbgaA::Pspr0276-lacZ; Tetr	This work
RW111	RW100 ΔbgaA::PpsaR-lacZ; Tetr	This work
RW112	D39 ΔbgaA::PpsaB-lacZ; Tetr	This work
RW113	RW100 ΔbgaA::PpsaB-lacZ; Tetr	This work
RW114	RW121 ΔbgaA::PpcpA-lacZ; Tetr	This work
RW120	RW103 with mariner insertion in psaR; Specr	This work
RW121	D39 ΔpsaCA::ermR; Ermr	This work
RW122	MP102 ΔbgaA::PpcpA-lacZ; Tetr	This work
RW131	D39 ΔbgaA::PpcpA-3-lacZ; Tetr	This work
RW132	D39 ΔbgaA::PpcpA-3a-lacZ; Tetr	This work
RW133	D39 ΔbgaA::PpcpA-4-lacZ; Tetr	This work
RW134	D39 ΔbgaA::PpcpA-5-lacZ; Tetr	This work
RW135	D39 ΔbgaA::PpcpA-rev3-lacZ; Tetr	This work
RW136	D39 ΔbgaA::PpcpA-rev4-lacZ; Tetr	This work
RW141	RW100 ΔbgaA::PpcpA-3-lacZ; Tetr	This work
RW142	RW100 ΔbgaA::PpcpA-3a-lacZ; Tetr	This work
RW143	RW100 ΔbgaA::PpcpA-4-lacZ; Tetr	This work
RW144	RW100 ΔbgaA::PpcpA-5-lacZ; Tetr	This work
RW145	RW100 ΔbgaA::PpcpA-rev3-lacZ; Tetr	This work
RW146	RW100 ΔbgaA::PpcpA-rev4-lacZ; Tetr	This work
E. coli EC1000	Kmr; MC1000 derivative carrying a single copy of the pWV01 repA gene in glgB	45
L. lactis NZ9000	MG1363 ΔpepN::nisRK	41
Plasmids
pR412T7	Specr; derivative of pR412 (49)	8
pORI13	Ermr; ori+repA−; promoterless lacZ, for single-copy chromosomal lacZ fusions.	70
pORI280	Ermr; ori+repA−; deletion derivative of pWV01; constitutive lacZ expression from P32 promoter	45
pPP2	Ampr Tetr; promoterless lacZ; for replacement of bgaA (spr0565) with promoter-lacZ fusions; derivative of pPP1	22
pNZ8048	Cmr; nisin-inducible PnisA	16
PNG8048E	Cmr Ermr; nisin-inducible PnisA, pNZ8048 derivative containing Ermr gene to facilitate cloning	Laboratory collection
pRW1	pORI13::prtA-lacZ	This work
pRW2	pPP2 Pspr0276	This work
pRW3	pPP2 PnrdD	This work
pRW4	pPP2 PpsaR	This work
pRW5	pPP2 PpcpA	This work
pRW6	pPP2 PpsaB	This work
pRW11	pPP2 PpcpA-3	This work
pRW12	pPP2 PpcpA-3a	This work
pRW13	pPP2 PpcpA-4	This work
pRW14	pPP2 PpcpA-5	This work
pRW15	pPP2 PpcpA-rev3	This work
pRW16	pPP2 PpcpA-rev4	This work
pRW20	pORI280 ΔpsaR	This work
pRW21	pNG8048E containing a 64-bp fragment comprising the PsaR binding site	This work
pRW22	pRW12 containing a point mutation (T→G, bp −186b) in the PsaR binding box	This work
pRW23	pRW12 containing a point mutation (A→G, bp −184b) in the PsaR binding box	This work
pRW24	pRW12 containing two point mutations (A→G, bp −184b; A→C, bp −185b) in the PsaR binding box	This work
pRW25	pNG8048E carrying psaR-strep downstream of PnisA	This work

^aErm^r, erythromycin resistance; Tet^r, tetracycline resistance; Cm^r, chloramphenicol resistance; Spec^r, spectimomycin resistance.

^bWhere the first base of the pcpA start codon (ATG) is +1.

TABLE 2.

Oligonucleotide primers used in this study

Primer	Nucleotide sequence (5′ to 3′)	Restriction site
PBMrIRPi	AGACCGGGGACTTATCAGCC
PBMrTn1	CTAGCGACGCCATCTATGTG
TMr_1	TGCATTTAATACTAGCGACGCCATCTATGTGTC
TMr_4	GGATCCATTCGCGTCAATTCGAGGGG
PpsaB-1-lacZ	CGGAATTCTTCCAAGTTTTTTACACTTG	EcoRI
PpsaB-2-lacZ	CGGGATCCATTGTTGGTCCATGGAGCAC	BamHI
pspr1945-1	CGGAATTCCCTTCAAATTTTAAGTCC	EcoRI
pspr1945-2	CGGGATCCGTTAATGATAATATTGTAG	BamHI
PpcpA_f_2	CGGAATTCTAATTTCTTTTTTAACCCAC	EcoRI
PpcpA_f_3	CGGAATTCGTGGGTTAATTTTCCTTGAC	EcoRI
PpcpA_f_3a	CGGAATTCTAAAAAAGAAATTAAAGTGG	EcoRI
PpcpA_f_4	CGGAATTCAATGTACTACTCTATTCTAC	EcoRI
PpcpA_rev_1	CGGGATCCAGGATTGGTTCATTAGGGAC	BamHI
PpcpA_rev_3	CGGGATCCGTCAAGGAAAATTAACCCAC	BamHI
PpcpA_rev_4	CGGGATCCCTAGTAGAATAGAGTAGTAC	BamHI
marR-lacZ1	CGGAATTCGCTATTTTCGTCATATCC	EcoRI
marR-lacZ2	CGGGATCCCATTTTAGATAGTCTTCTTTG	BamHI
marR-del1	TGCTCTAGACAATTGCCCACCAGTCCCG	XbaI
marR-del2	TTCTTTGTTTGGGGTCATTC
marR-del3	GACCCCAAACAAAGAAGTCGAGAAAATCAACTAAT
marR-del4	GAAGATCTCTTTTGTCAGCTGAACGA	BglII
PprtA-1	CGGAATTCTAGCGCTGATATTTCATC
PprtA-2	CGGGATCCGTTAGTGACAATACTGTG
spr0561-1	CGGAATTCGATACGGGAGAGGTAAGTG	EcoRI
spr0561-2	CGGGATCCGAAATTGTCTCATCACCTC	BamHI
PSpr183-1.2	CGGAATTCCAACCTAAGGTGATTGTGG	EcoRI
PSpr183-2.2	CGGGATCCGAATTTCTGTAATAATTCGC	BamHI
Pspr0276-1	CGGAATTCGAATTAGAATTATTGTGAG	BamHI
Pspr0276-2	CGGGATCCGAGCAGCAACAGCACCACC	XbaI
Ery-for	TAACGATTATGCCGATAACT
Ery-rev	GCATGCATCGATTAGATCTC
psaCA-KO-1	AAATCCTTACACCGAATTGC
psaCA-KO-2	GAGATCTAATCGATGCATGCTTCTGCAATCATAGGTCACCTCC
psaCA-KO-3	AGTTATCGGCATAATCGTTAGACAAGATTGCTGAAGGATTGG
psaCA-KO-4	TGCGCGTGCTAATAGGTGCC
P-pcpA-box1	CATGTCTTTTTTAACACGGGTTAAAAAAGAAATTAAAGTGGGTTAATTTTCCTTGACTTAAAATTTAA
P-pcpA-box2	CTAGTTAAATTTTAAGTCAAGGAAAATTAACCCACTTTAATTTCTTTTTTAACCCGTGTTAAAAAAGA
PcpA_mut1.1	GATTTCTTTTTTAACCCGTG
PcpA_mut1.2	AAAGTGGGTTAATTTTCCTTG
PpcpA_mut2.1	GTAATTTCTTTTTTAACCCGTG
PpcpA_mut2.2	AGTGGGTTAATTTTCCTTGAC
PpcpA_mut3.1	GCAATTTCTTTTTTAACCCGTG
PpcpA_mut3.2	AGTGGGTTAATTTTCCTTGAC
Spr1480OX-2	TGCTCTAGATTAGTTGATTTTCTCGACATAG	XbaI
spr1480-OX-new	CGAGCCATCATGAAACATCATCATCATCATCATACCCCAAACAAAGAAGACTATC	RcaI
Spr1480-Ctermstrep_OX	TGCTCTAGATTATTTTTCAAATTGTGGATGGCTCCAAGCGCTGTTGATTTTCTCGACATAGAGTTG	XbaI
Spr1480OX-1	CGAGCCATCATGACCCCAAACAAAGAAGAC	RcaI

Construction of transcriptional lacZ fusions.

Ectopic lacZ fusions to the pcpA (spr1945), psaR (spr1480), psaB (spr1492), nrdD (spr0183), and spr0276 promoters were made in pPP2 with primer pairs pspr1945-1/pspr1945-2, marR-lacZ1/marR-lacZ2, PpsaB-1-lacZ/PpsaB-2-lacZ, Pspr183-1.2/Pspr183-2.2, and Pspr0267-1/Pspr0276-2, yielding plasmids pRW5, pRW4, pRW6, pRW3, and pRW2. A chromosomal lacZ fusion to the 3′ end of prtA (spr0561) was made with primer pair spr0561-1/spr0561-2 in plasmid pORI13, leading to plasmid pRW1. In all cases Escherichia coli EC1000 was used as the cloning host. The lacZ fusion constructs were introduced into wild-type D39 and D39 ΔpsaR (RW100) in the case of pPP2 (integration via double crossover in the bgaA gene) and into D39nisRK and its isogenic psaR mutant (RW101) in the case of pORI13 (integration by single crossover), giving strains RW102 to RW113. The PpcpA-lacZ fusion was also introduced into the psaCA (RW121) and czcD (MP102) deletion strains, giving strains RW114 and RW122. All plasmid constructs were checked by sequencing, and new loci created with these plasmids were verified by PCR.

Random mutagenesis screen.

Random mutagenesis using the Himar1 MarC9 mariner transposon was performed essentially as described previously (42, 49). pR412-T7 (8), a derivative of pR412, was used as the source of the spec mariner transposon. Mutated R6 chromosomal DNA was transformed into strain R6, yielding a mutant library of approximately 20,000 CFU. Chromosomal DNA of this library was used to perform random mutagenesis in strain D39 ΔbgaA::PpcpA-lacZ. Mutants were selected on GM17 with 1% sheep blood, 0.006% X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside), 1.5 μg/ml tetracycline, and 130 μg/ml spectinomycin. Transposon insertion sites in mutants showing derepression of expression of the PpcpA-lacZ transcriptional fusion were identified by the direct sequencing method as described previously (34), using primer PBMrTn1 with PBMrIRPi (nested).

Construction of deletion strains.

An in-frame marker-free deletion of psaR (spr1480) was constructed with plasmid pORI280 essentially as described previously (37), using primer pairs marR-del1/marR-del2 and marR-del3/marR-del4 and E. coli EC1000 as the cloning host. The mutant was checked by PCR and DNA sequencing.

A deletion strain of psaCA was made with primer pairs psaCA-KO-1/psaCA-KO-2 and psaCA-KO-1/psaCA-KO-4 by overlap extension PCR (73) and allelic replacement with an erythromycin resistance cassette, yielding strain RW121.

Construction of pcpA promoter subclones in pPP2.

The following promoter subclones of the pcpA promoter were made in pPP2 (primer pairs are in parentheses): PpcpA-3 (PpcpA_f_2 and PpcpA_rev_1), PpcpA-3a (PpcpA_f_3a and PpcpA_rev_1), PpcpA-4 (PpcpA_f_3 and PpcpA_rev_1), PpcpA-5 (PpcpA_f_4 and PpcpA_rev_1), PpcpA-rev3 (spr1945-1and PpcpA_rev_3), and PpcpA-rev4 (spr1945-1 and PpcpA_rev_4). All fragments were cloned as EcoRI/BamHI fragments in the same sites of pPP2, yielding plasmids pRW11 to pRW16. The constructs were sequenced and introduced into strains D39 and D39ΔpsaR (RW100), giving strains RW131 to RW136 and RW141 to RW146.

Construction of point mutations in the PsaR binding site in PpcpA.

A pNG8048E derivative containing a 64-bp fragment of the pcpA promoter comprising the PsaR binding site was constructed by annealing oligonucleotides P-pcpA-box1/P-pcpA-box2 and cloning them into NcoI/XbaI-digested pNG8048E, giving plasmid pRW21. This plasmid was used as a template for a PCR with Phusion DNA polymerase using the primer pairs PcpA_mut1.1/PcpA_mut1.2, PpcpA_mut2.1/PpcpA_mut2.2, and PpcpA_mut3.1/PpcpA_mut3.2, giving rise to plasmids pRW22 and pRW23, each with one point mutation in the predicted PsaR binding site, and pRW24, with two point mutations in the predicted PsaR binding site, respectively. The constructs were checked by DNA sequencing.

Microarray analyses.

For microarray analysis, the transcriptome of wild-type D39 grown in four biological replicates in GM17 was compared to the transcriptome of the same strain grown in four biological replicates in GM17 plus 0.25 mM ZnSO₄. Growth, RNA isolation, labeling, hybridization, and analysis of slides were done essentially as described previously (38). Since the analysis was performed by interslide comparisons (i.e., the wild-type transcriptome analyzed with four slides was compared to the wild-type transcriptome in GM17 plus Zn²⁺ analyzed with four different slides), scaled signals from PostPrep of the slides hybridized to the GM17 samples were pairwise compared with the scaled signals of the slides hybridized to the GM17-Zn²⁺ samples. The resultant table with the scaled (and normalized) signals was used as input for the CyberT variant for statistical analysis of control versus experimental data, using a local running copy of the CyberT algorithm for paired data (http://molgen.biol.rug.nl/cgi-bin/cybert/CyberT-8.0.form.pl?DATATYPE=CE). Genes were considered differentially expressed when the Bayesian P value was <0.001 and the false discovery rate was <0.01, unless otherwise stated (80).

Overexpression and purification of Strep-tagged PsaR.

For the overexpression of a C-terminally Strep-tagged variant of PsaR, psaR was amplified from D39 chromosomal DNA using primers Spr1480OX-1 and Spr1480-Ctermstrep_OX. The resulting PCR product was digested with RcaI/XbaI and cloned into the NcoI/XbaI sites of pNG8048E, yielding plasmid pRW25. Overexpression in Lactococcus lactis NZ9000 was done essentially as described previously (41). Purification of PsaR-Strep from L. lactis was performed using the Streptactin column from IBA according to the supplier's instructions. Buffers without EDTA were used, and the purified protein was stored at a concentration of 0.15 mg/ml in the elution buffer (100 mM Tris-HCl [pH 8], 150 mM NaCl, 2.5 mM desthiobiotin, 1 mM β-mercaptoethanol) with 10% glycerol at −80°C.

EMSAs.

Electrophoretic mobility shift assays (EMSAs) were performed with [γ-³³P]ATP-labeled probes in buffer containing 20 mM Tris-HCl (pH 8.0), 5 mM MgCl₂, 0.1 mM dithiothreitol, 8.7% (wt/vol) glycerol, 62.5 mM KCl, 25 μg/ml bovine serum albumin, 25 μg/ml poly(dI-dC), and 3000 cpm of [γ-³³P]ATP-labeled PCR product. Various metal ions were added in concentrations as specified in Results. As probes, PCR products comprising the promoter regions of psaB (primers PpsaB-1-lacZ/PpsaB-2-lacZ), pcpA (PpcpA_f_2/PpcpA_rev_1), and prtA (PprtA-1/PprtA-2) were used. As a control, a PCR fragment of the pcpA promoter without the intact PsaR operator was used (PpcpA_f_4/PpcpA_rev_1). Reaction mixtures were incubated at 37°C for 10 min before loading on gels. Gels were run in 0.44 M Tris-borate buffer (pH 8.3) at 100 V for 90 min.

Measurement of concentrations of Mn²⁺ and Zn²⁺ in growth media.

Using atomic absorption spectroscopy on a Vista AX-CCD simultaneous ICP-AES spectrometer, the concentrations of Mn²⁺ and Zn²⁺ in GM17 and GM17chel were determined at wavelengths (nm) of 257.610/259.372 (Mn) and 202.548/213.857 (Zn).

Microarray accession number.

Microarray data have been deposited to the Gene Expression Omnibus (GEO) and can be accessed via http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE11438. In addition, slide images and raw and processed data are available at http://molgen.biol.rug.nl/publication/zinc_data/.

RESULTS

Identification of Zn²⁺-regulated genes in S. pneumoniae.

To identify genes that are regulated by a high level of Zn²⁺ in S. pneumoniae, we compared the transcriptome of wild-type D39 grown in GM17 with that of the same strain grown in GM17 with 0.25 mM ZnSO₄ (Table 3). With this concentration of Zn²⁺, growth is not affected (39). The most upregulated genes were pcpA, encoding a choline binding protein (69) involved in virulence (31), and Zn²⁺ resistance gene czcD (39). Also upregulated were the serine protease gene prtA and the Mn²⁺ ABC transporter genes psaBC, which are both involved in virulence (6, 7, 32, 50), and an operon (SP0202 to SP0207) encoding, among other proteins, the NrdDG ribonucleoside triphosphate reductase, which is involved in synthesis of deoxyribonucleoside triphosphates (33, 78). An operon consisting of genes involved in cellobiose metabolism (SP0303 to SP0310) (52) was strongly downregulated. Thus, expression of genes with a variety of functions is affected by growth of S. pneumoniae in medium with an elevated Zn²⁺ concentration.

TABLE 3.

Summary of transcriptome comparison of S. pneumoniae strain D39 grown in GM17 and in GM17 with addition of 0.25 mM Zn²⁺

TIGR4 locus tag	Function (TIGR annotation)	Ratioa
SP0202	Anaerobic ribonucleoside triphosphate reductase NrdD	2.5
SP0204	Predicted acetyltransferase, GNAT family	2.4
SP0205	Anaerobic ribonucleoside triphosphate reductase activating protein NrdG	1.8b
SP0206	Hypothetical protein; predicted uridine kinase	1.8b
SP0207	Hypothetical protein; predicted uridine kinase	2.2
SP0303	6-Phospho-β-glucosidase BglA	−7.6
SP0305	Phosphotransferase system cellobiose-specific component IIB	−4.1
SP0306	Putative transcriptional regulator; possible antiterminator BglG	−14.8
SP0307	Phosphotransferase system, IIA component	−6.3b
SP0308	Phosphotransferase system cellobiose-specific component IIA	−3.0
SP0309	Hypothetical protein	−7.6
SP0310	Phosphotransferase system cellobiose-specific component IIC	−6.0
SP0338	Putative ATP-dependent clp protease, ATP binding subunit ClpL	−2.0
SP0515	Heat-inducible transcription repressor HrcA	−2.0
SP0516	Heat shock protein GrpE	−1.6
SP0517	Chaperone protein DnaK (heat shock protein 70)	−1.4
SP0518	Hypothetical protein	−2.1
SP0519	Chaperone protein DnaJ	−1.4
SP0640	Hypothetical protein	1.5b
SP0641	Cell wall-associated serine proteinase precursor PrtA	2.7
SP0645	Putative phosphotransferase system IIA component	2.8
SP0646	Phosphotransferase system, IIB component, putative	2.4
SP0879	Hypothetical protein	−2.1
SP1648	Manganese (and/or zinc) ABC transporter, ATP binding protein PsaB	2.3
SP1649c	Manganese (and/or zinc) ABC transporter, permease protein PsaC	2.6
SP1762	Hypothetical protein	2.7
SP1855	Alcohol dehydrogenase, zinc-containing AdhB	1.7
SP1856	Transcriptional regulator, MerR family	2.0b
SP1857	Cation efflux system protein CzcD	7.2
SP1935	Hypothetical protein	2.8
SP2136	Choline binding protein; surface protein PcpA	8.5

^aRatios of >2.0 or <−2.0 (wild-type D39 compared to wild-type D39 plus 0.25 mM Zn²⁺) are shown; in some cases neighboring genes with lower fold changes are also indicated.

^bRatio with a false discovery rate of >0.01 but <0.1.

^cFor SP1650 (psaA), the number of observations (replicates) was too low and hence no significance was obtained.

Zn²⁺-dependent expression of pcpA is mediated by PsaR.

To investigate in more detail the transcriptional regulation of pcpA, the most Zn²⁺-induced gene found in the microarray experiment, a transcriptional lacZ fusion to the pcpA promoter was constructed by use of plasmid pPP2, which integrates into the bgaA locus (22). Using strain D39 containing this PpcpA-lacZ transcriptional fusion, it was demonstrated that only elevated Zn²⁺ concentrations lead to high expression from the pcpA promoter, although also some weak induction could be seen for Co²⁺ and Fe²⁺ (Table 4).

TABLE 4.

Expression of PpcpA specifically in the presence of various metal cations^a

Addition to GM17 (mM)	Mean (SD) β-galactosidase activity (Miller units)
None	4 (1)
Zn2+ (0.2)	54 (12)
Zn2+ (0.4)	388 (45)
Cu2+ (0.05)	3 (1)
Cu2+ (0.1)	3 (1)
Co2+ (0.05)	8 (3)
Co2+ (0.1)	16 (3)
Ni2+ (0.1)	4 (1)
Ni2+ (0.4)	5 (2)
Fe2+ (0.1)	9 (2)
Fe2+ (0.4)	21 (6)
Mg2+ (1.0)	4 (1)
Mg2+ (10)	2 (1)

^aβ-Galactosidase of a PpcpA-lacZ transcriptional fusion was measured in the wild-type strain D39 (strain RW102) grown in GM17 with the indicated metal ions. For Zn²⁺, Ni²⁺, Co²⁺, and Cu²⁺ the concentrations used have similar effects on growth (see also reference 39). Values are from three experiments.

To identify the factor repressing transcription of pcpA under normal growth conditions, strain D39 PpcpA-lacZ (RW102) was randomly mutagenized with the mariner transposon (42, 49) and blue colonies were screened for on GM17 agar plates with X-Gal but without Zn²⁺ supplementation. Among 7,200 CFU, one blue clone was found, containing a transposon insertion in the gene encoding the MarR family transcriptional regulator PsaR (spr1480, SPD_1450, SP1638). The insertion in psaR gave rise to high expression of the PpcpA-lacZ fusion, which was independent of Zn²⁺ (data not shown). We constructed a markerless deletion mutant of psaR and found that in this mutant expression of PpcpA-lacZ is highly derepressed independent of Zn²⁺ (Table 5), suggesting that PsaR is responsible for Zn²⁺-dependent derepression of pcpA expression.

TABLE 5.

β-Galactosidase activities of the wild-type D39 and the ΔpsaR, ΔpsaCA, and ΔczcD strains, all containing the PpcpA-lacZ transcriptional fusion, grown in GM17 and GM17chel supplemented with metal ions^a

Medium and addition(s) (mM)	Mean (SD) β-galactosidase activity (Miller units)b in:
	Wild type	ΔpsaR mutant	ΔpsaCA mutant	ΔczcD mutant
GM17
None	6 (3)	1,204 (67)	594 (55)	18 (5)
Zn2+ (0.1)	17 (4)	ND	ND	69 (12)
Zn2+ (0.2)	66 (21)	1,102 (277)	ND	ND
Zn2+ (0.4)	417 (113)	1,189 (115)	446 (37)	ND
Zn2+ (0.4) + Mn2+ (0.01)	33 (12)	1,119 (113)	237 (22)	ND
Zn2+ (0.4) + Mn2+ (0.05)	11 (4)	1,219 (203)	7 (4)	ND
Zn2+ (0.4) + Fe2+ (0.05)	269 (22)	ND	ND	ND
GM17chel
None	315 (47)	1,217 (222)	ND	ND
Zn2+ (0.2)	424 (35)	1,260 (102)	ND	ND
Mn2+ (0.05)	6 (2)	1,315 (78)	ND	ND

^aThe wild-type D39 and the ΔpsaR, ΔpsaCA, and ΔczcD strains are strains RW102, RW107, RW114, and RW122, respectively.

^bValues are from three experiments. ND, not determined.

PsaR regulates prtA and psaBCA in a Zn²⁺-dependent way.

To test whether PsaR is also responsible for the Zn²⁺-dependent expression of the other genes identified in the microarray analysis, transcriptional lacZ fusions to prtA, PpsaB, and PnrdD (Pspr0183) were constructed and introduced in both the wild type and the psaR mutant. As expected, in the wild type expression of prtA and PpsaB increased upon addition of Zn²⁺ to the GM17 growth medium (Tables 6 and 7). In the psaR mutant expression was derepressed in GM17, showing that PsaR mediates the Zn²⁺-dependent expression of prtA and psaBCA as well. Transcription from PnrdD was two- to threefold higher in cells grown in GM17 with Zn²⁺ compared to GM17, but this was not affected by the psaR mutation (data not shown). We also tested the effect of metal ions on the expression of psaR itself. A PpsaR-lacZ transcriptional fusion was highly expressed in GM17 but was not influenced by Zn²⁺ and Mn²⁺ or by the psaR mutation (data not shown).

TABLE 6.

β-Galactosidase activities of the wild-type D39 and the ΔpsaR strain, containing the prtA-lacZ transcriptional fusion, grown in GM17 and GM17chel supplemented with metal ions^a

Medium and addition(s) (mM)	Mean (SD) β-galactosidase activity (Miller units)b in:
	Wild type	ΔpsaR mutant
GM17
None	0.2 (0.06)	2.2 (0.3)
Zn2+ (0.2)	0.6 (0.17)	2.3 (0.5)
Zn2+ (0.4)	1.2 (0.05)	2.1 (0.4)
Zn2+ (0.4) + Mn2+ (0.01)	0.5 (0.09)	2.1 (0.2)
Zn2+ (0.4) + Mn2+ (0.05)	0.2 (0.08)	2.1 (0.2)
GM17chel
None	0.5 (0.03)	2.2 (0.1)
Zn2+ (0.2)	0.9 (0.05)	2.0 (0.4)
Mn2+ (0.05)	0.2 (0.04)	2.1 (0.4)

^aThe wild-type D39 and the ΔpsaR strain are strains RW104 and RW109, respectively. lacZ was fused to the 3′ end of prtA on the native chromosomal location, using plasmid pORI13. This might explain the much lower β-galactosidase activity compared to the values for the lacZ fusions with PpcpA and PpsaB (Tables 5 and 7).

^bValues are from three experiments.

TABLE 7.

β-Galactosidase activities of the wild-type D39 and the ΔpsaR strain, containing the PpsaB-lacZ transcriptional fusion, grown in GM17 and GM17chel supplemented with metal ions^a

Medium and addition(s) (mM)	Mean (SD) β-galactosidase activity (Miller units)b in:
	Wild type	ΔpsaR mutant
GM17
None	134 (34)	1,167 (121)
Zn2+ (0.2)	448 (110)	970 (78)
Zn2+ (0.4)	760 (124)	1,089 (98)
Zn2+ (0.4) + Mn2+ (0.01)	651 (18)	1,205 (144)
Zn2+ (0.4) + Mn2+ (0.05)	142 (11)	1,072 (262)
GM17chel
None	583 (41)	1,107 (220)
Zn2+ (0.2)	901 (33)	1,180 (102)
Mn2+ (0.05)	111 (27)	1,023 (45)

^aThe wild-type D39 and the ΔpsaR strain are strains RW112 and RW113, respectively.

^bValues are from three experiments.

The expression of the cellobiose utilization operon SP303 to SP310 (spr0276 to spr0282, R6 annotation) was further analyzed using a Pspr0276-lacZ transcriptional fusion. Surprisingly, in GM17 the expression was hardly measurable, and no effect of Zn²⁺ was observed (data not shown). However, in the absence of glucose as well as in the presence of cellobiose, Pspr0276 was highly expressed (data not shown). It is not immediately clear what the reason is for the downregulation of this operon in the transcriptome analysis.

Regulation of PpsaB, PpcpA, and prtA depends on the balance between Mn²⁺ and Zn²⁺.

In an earlier study, expression of pcpA and psaBCA was demonstrated to be repressed by PsaR in the presence of Mn²⁺ (31). To specify how Mn²⁺ and Zn²⁺ influence PsaR activity, we analyzed the expression of transcriptional lacZ fusions to PpcpA, PpsaB, and prtA in the presence of various concentrations of both Mn²⁺ and Zn²⁺ in GM17 medium and in GM17 medium treated with the metal ion chelator Chelex 100 resin (GM17chel). GM17 contains 2.0 μM Mn²⁺ and 8.1 μM Zn²⁺ and GM17chel contains 0.02 μM Mn²⁺ and 0.00 μM Zn²⁺, showing that the Chelex treatment effectively removed these cations. In GM17, upregulation of the expression of all three lacZ fusions in the presence of a high concentration of Zn²⁺ was nullified by the addition of Mn²⁺ when added in a concentration of 0.01 to 0.05 mM, which is about 10 to 40 times lower than the concentration of Zn²⁺ (Tables 5, 6, and 7). This repressive effect was specific for Mn²⁺, since 0.05 mM Ni²⁺, Cu²⁺, and Co²⁺ had no effect on the Zn²⁺-dependent expression of the PpcpA-lacZ transcriptional fusion (data not shown); for Fe²⁺, there was only a weak repressive effect (Table 5). In GM17chel, expression of all three lacZ fusions was derepressed compared to that in GM17 (Tables 5, 6, and 7), which was expected because of the much lower concentration of Mn²⁺ after Chelex treatment. Addition of Zn²⁺ increased the derepression even more, while Mn²⁺ led to repression of expression again, which is in agreement with the observations made in GM17 (Tables 5, 6, and 7).

To investigate whether the opposite effects of Zn²⁺ and Mn²⁺ are the result of the competition for uptake of these cations, expression of the PpcpA-lacZ fusion in czcD and psaCA deletion mutants was measured. In a czcD deletion mutant, which is, as a consequence of impaired Zn²⁺ efflux, likely to accumulate higher intracellular levels of this metal ion (39), expression of PpcpA-lacZ in both GM17 and GM17 with 0.1 mM Zn²⁺ (the highest possible concentration for the czcD deletion mutant) was higher than that in the wild type (Table 5). In a psaCA deletion mutant, which is impaired in uptake of Mn²⁺ into the cell (50) the expression of PpcpA-lacZ was also highly derepressed, However, addition of Mn²⁺, albeit at a higher concentration than with the wild type, still led to repression (Table 5). These results suggest that the observed regulatory effects on expression of pcpA, prtA, and psaBCA that are induced by Mn²⁺ and Zn²⁺ converge at the level of transcriptional regulation by PsaR.

Identification of a PsaR operator in the promoters of pcpA, prtA, and psaB.

Using Gibbs Motif Sampler (77), a palindromic sequence (Fig. 1A and B), located just upstream of (PprtA and PpcpA) or overlapping with (PpsaB) the predicted core promoter regions, that might serve as the PsaR operator was uncovered.

FIG. 1.

Identification of a putative PsaR operator. (A) Weight matrix of the identified PsaR operator as present in the promoter regions of pcpA, prtA, and psaB. (B) Positions of the PsaR operator (shaded) in the promoter regions of pcpA, prtA, and psaB. Putative core promoter sequences are in bold. The ribosome binding sites are in bold and underlined. Start codons are in italic.

To dissect the promoter of pcpA experimentally, we performed a promoter subcloning experiment where the pcpA promoter was truncated from the 5′ end and fused to lacZ in the reporter plasmid pPP2 (Fig. 2). Expression of a promoter fragment truncated upstream of the predicted operator (PpcpA-3) was Zn²⁺ dependent, but as expected, deletion of half of the identified operator (PpcpA-4) led to fully derepressed, Zn²⁺-independent expression (Fig. 2). In the presence of 0.05 mM Mn²⁺, the β-galactosidase activity of PpcpA-4 in wild-type D39 was 1,197 ± 78 Miller units, which is similar to the values for PpcpA-4 in GM17 and GM17 with Zn²⁺ (Fig. 2), indicating that the effects of Zn²⁺ and Mn²⁺ are mediated by the same PsaR operator site. To determine if the operator sequence was fully identified, a truncation of the promoter region 8 bp upstream of the operator sequence (PpcpA-3a) was constructed, which gave rise to full derepression, suggesting that additional bases 5′ to the operator are also important for PsaR-mediated repression of the pcpA promoter. For subclone PpcpA-5, expression was close to zero under all conditions. Deletion of the same region as in PpcpA-5 but now from the 3′ side (PpcpA-rev3 versus PpcpA-rev4; strains RW135, RW136, and RW145, RW146) confirmed that promoter activity locates exclusively to this area (data not shown). This demonstrates that the core promoter sequence is located in the region between the 5′ base pair positions of subclones PpcpA-4 and PpcpA-5 (Fig. 2 and 1B).

FIG. 2.

Subcloning of PpcpA. A schematic overview of the PpcpA truncations is shown. Numbers indicate the positions of the truncations, which were fused to lacZ, relative to the putative pcpA start site (+1). The flag indicates the position of the core promoter, and the oval indicates the putative PsaR operator. Gray-shaded areas indicate regions of similarity with PprtA (89% identity for the short stretch and 82% identity for the long stretch). The table on the right gives β-galactosidase activities (Miller units) of the promoter truncations in wild-type D39 (wt, strains RW131 to RW134) and the ΔpsaR mutant (psaR, strains RW141 to RW144) grown in GM17 and in GM17 plus 0.5 mM Zn²⁺. Standard deviations of three measurements are given in parentheses.

To further show that the predicted PsaR operator is functional, a 64-bp DNA fragment of the pcpA promoter comprising the PsaR operator was put into plasmid pNG8048E, which replicates in S. pneumoniae (37). Subsequently, several point mutations in the first half of the motif were introduced (Table 8). By putting the wild-type construct into D39 containing the PpcpA-lacZ transcriptional fusion, transcription from PpcpA was strongly derepressed, showing that this 64-bp stretch of DNA titrates away the repressive effect of PsaR on the expression of pcpA (Table 8). However, with the constructs containing the mutated PsaR boxes, this derepressive effect was not present (Table 8). This shows clearly that the bases in the predicted PsaR binding box are required for PsaR-dependent repression of pcpA. The entire S. pneumoniae R6, D39, and TIGR4 sequences were searched with a weight matrix of the PsaR operator sequence (Fig. 1A) using Genome2D (5), but the motif was not found in additional promoter regions (data not shown). In conclusion, a PsaR regulatory element in the promoters of pcpA, psaB, and prtA was identified.

TABLE 8.

Mutational analysis of the PsaR operator

PsaR box	Sequence (5′→3′)a	Mean (SD) β-galactosidase activity (Miller units) for the following strain and mediumb:
		Wild type		ΔpsaR mutant
		GM17	GM17 + Zn2+ (0.4 mM)	GM17	GM17 + Zn2+ (0.4 mM)
Wild type	AAATTAAAGTGGGTTAATTT	733 (123)	853 (122)	1,288 (170)	1,331 (167)
mut 1	AAATGAAAGTGGGTTAATTT	74 (15)	376 (78)	1,160 (135)	1,271 (122)
mut 2	AAATTAGAGTGGGTTAATTT	121 (25)	416 (75)	1,300 (101)	1,278 (69)
mut 3	AAATTCGAGTGGGTTAATTT	58 (12)	464 (16)	1,389 (111)	1,378 (178)

^aSequences of wild-type and mutant PsaR boxes of the pcpA promoter; point mutations are in bold.

^bβ-Galactosidase activities of wild-type D39 and the ΔpsaR mutant harboring the PpcpA-lacZ transcriptional fusion (strains RW102 and RW107) containing plasmid pRW21 (wild-type PsaR box), pRW22 (PsaR box mut1), pRW23 (PsaR box mut2), or pRW24 (PsaR box mut3).

Binding of PsaR-Strep to the pcpA, psaB, and prtA promoters in the presence of Mn²⁺ is counteracted by Zn²⁺.

To find out whether the observed Zn²⁺- and Mn²⁺-dependent effects on expression of the PsaR targets are caused by direct modulation of the DNA binding activity of the PsaR protein, EMSAs were performed with purified Strep-tagged PsaR (PsaR-Strep). PsaR-Strep alone did not shift the promoter regions of pcpA, psaB, and prtA (Fig. 3A to C, lanes 2). However, in the presence of Mn²⁺, PsaR-Strep did bind to the promoter regions of pcpA, psaB, and prtA (Fig. 3A to C, lanes 3). PsaR-Strep did not bind under any condition to a truncated pcpA promoter lacking the PsaR binding box (Fig. 3D). Besides Mn²⁺, 0.05 mM Co²⁺ was also able to stimulate the binding of PsaR-Strep to the promoter fragments, whereas 0.05 mM Fe²⁺, Cu²⁺, Ni²⁺, and Zn²⁺ did not (data not shown). These data show that the PsaR-Strep-DNA interaction was specific and indicates that PsaR directly functions as a Mn²⁺-dependent repressor of its target genes. Based on the lacZ expression studies, we hypothesized that Zn²⁺ should somehow impair PsaR binding to fulfill its function as an Mn²⁺-dependent repressor. Therefore, experiments addressing the influence of Zn²⁺ on the in vitro PsaR-Strep-DNA interaction in the presence of Mn²⁺ were performed (Fig. 3A, B, and C, lanes 4 to 8). These demonstrated that the stimulatory effect of Mn²⁺ on the binding of PsaR-Strep to all three promoter fragments was counteracted by the addition of Zn²⁺. There was also a weaker counteracting effect of Cu²⁺ for the pcpA and prtA promoters (Fig. 3A and C, lanes 8). Thus, Mn²⁺ stimulates PsaR binding to the operators in the pcpA, psaB, and prtA promoters, whereas in the presence of Zn²⁺, PsaR binding is abolished, indicating that Mn²⁺ and Zn²⁺ exert their regulatory effects on pcpA, prtA, and psaBCA expression directly through PsaR.

FIG. 3.

In vitro interaction of PsaR-Strep with the pcpA (A), psaB (B), and prtA (C) promoter regions and with a truncated PpcpA fragment lacking the PsaR operator (D). Purified PsaR-Strep was added at concentrations of 5 nM (PpcpA), 11 nM (PprtA), and 25 nM (PpsaB). Metal ions were added as indicated above the lanes at a concentration of 50 μM. X, free probe. The horizontal bar above lanes 3 to 8 indicates the presence of Mn²⁺, and the horizontal bar above lanes 2 to 8 indicates the presence of PsaR-Strep. Arrows indicate the positions of the shifted probes, and asterisks indicate the position of the free probe. The presence of weaker bands which run higher than the free probe in the gels is a phenomenon that has also been seen by others in similar experiments. These bands may represent unspecific PCR products or single-stranded DNA (1, 15).

DISCUSSION

In this study, we analyzed the transcriptome change of the human pathogen S. pneumoniae in response to a high level of Zn²⁺. Expression of several genes and operons with diverse functions was affected by Zn²⁺, including pcpA, prtA, and psaBCA. The observed Zn²⁺-dependent expression of these virulence genes was shown to be directly mediated by the Mn²⁺-responsive repressor PsaR (31). We further demonstrate that this is caused by Mn²⁺-dependent binding of PsaR to and Zn²⁺-dependent release from the promoters of these genes. Thus, these data represent an intriguing insight in the opposite regulatory effects of two metal cations on the expression of a set of virulence genes, mediated by a single transcriptional repressor.

The concentrations as well as the ratio of Mn²⁺ and Zn²⁺ may vary greatly between different sites in the human body. For example, in lung tissue the total concentration of Mn²⁺ is approximately 0.2 μg/g (wet weight) (3.6 μM) and that of Zn²⁺ is 15 μg/g (wet weight) (229 μM), whereas in the blood serum concentrations of Mn²⁺ and Zn²⁺ are 0.5 ng/ml (9 nM) and 1.0 μg/ml (15.3 μM), respectively (81). Control of availability of both ions is of importance to the host, as a recent study showed that the human immune system employs chelation of Mn²⁺ and Zn²⁺ by calprotectin as a way to inhibit bacterial growth in tissue abscesses (13). On the other hand, sufficiently high levels of Zn²⁺ are required for proper functioning of the immune system (72). Thus, it is likely that the concentrations of Mn²⁺ and Zn²⁺ fluctuate greatly in the environment, which will lead to varying concentrations of these metal ions in the cytoplasm of S. pneumoniae. However, virulence studies that have been carried out so far with respect to pcpA, prtA, and psaBCA do not point to a specific site where the proteins encoded by these genes are needed (7, 26, 31, 47, 50, 51, 62). Interestingly, pcpA is also regulated by the nutritional regulator CodY (26), meaning that the concentrations of both metal ions as well as amino acids affect pcpA expression.

Homologs of PsaR in other organisms seem to have slightly different functions. In Streptococcus gordonii, ScaR is an Mn²⁺-dependent repressor of the Sca (Mn²⁺) permease (29). In Streptococcus pyogenes, MtsR regulates the Mn²⁺-specific ABC transporter MtsABC in response to Mn²⁺, while the heme-specific HtsABC transporter is repressed by MtsR in response to Fe³⁺ (23). The Streptococcus mutans SloR regulates several genes involved in biofilm formation, genetic competence, oxidative stress tolerance, and adherence in response to Mn²⁺ and, to a lesser extent, Fe³⁺ (35, 63, 67, 74).

The EMSAs performed in this study are in line with the transcriptional data and suggest that a direct effect of Zn²⁺ and Mn²⁺ on the PsaR-promoter interaction causes the observed transcriptional effects. However, in the EMSAs Mn²⁺ does not overcome the Zn²⁺ effect at equimolar concentration, whereas in vivo, only low (but repressive) concentrations of Mn²⁺ are counteracted by Zn²⁺. This might be because of different intracellular concentrations/availabilities of these metal ions compared to the extracellular concentrations and indicates that in vivo Mn²⁺ is the principal effector. It will be interesting to know why Zn²⁺ and Mn²⁺ have these opposite effects on the DNA binding properties and activity of PsaR.

Clues about this could come from recent structural studies on DtxR from Corynebacterium diphtheriae (14) and MntR, a DtxR family protein from Bacillus subtilis that responds to Mn²⁺ (65). Both DtxR and MntR contain two metal binding sites per monomer: a low-affinity site and a high-affinity site (14, 21). MntR binds metal ions with affinities that roughly follow the Irving-Williams series, where Mn²⁺ displays the lowest affinity for MntR and Zn²⁺ the highest (21). As MntR has a much higher affinity for Zn²⁺ than Mn²⁺ but only very poorly activates DNA binding of MntR, the specificity of MntR is not correlated with the metal binding affinity (21). The conformation of Mn²⁺-bound MntR differs from the Zn²⁺-bound conformation with respect to the occupancy of the metal binding sites: Mn²⁺ binds to two sites, whereas only one Zn²⁺ ion binds to MntR, which does not allow binding of a second one (36). Metal binding at the second site is proposed to be required for DNA binding, as it promotes a disorder-to-order transition of MntR structure (17). PsaR shares 25% and 15% sequence identity with DtxR and MntR, respectively. Moreover, sequence alignment shows that most residues that constitute the metal ion binding sites in DtxR and MntR are conserved in PsaR (data not shown). Therefore, it might be that Zn²⁺ prevents Mn²⁺ binding to PsaR, rendering PsaR in a monomeric or destabilized state, and in this way counteracts Mn²⁺-induced DNA binding and transcriptional repression.

Interference with the effect of one metal ion on a metal-sensory protein by another metal ion has been reported recently for CzrA in B. subtilis (24). CzrA normally is activated for DNA release in the presence of Zn²⁺, but Cu²⁺ inhibits the Zn²⁺-induced allostery, since in vitro the protein preferentially binds Cu²⁺. However, these effects are not seen in vivo. High levels of Cu²⁺ in the growth medium induces the Fur regulon in B. subtilis (56). Thus, opposite effects of metals on regulation of gene expression seems to occur with other classes of metalloregulatory proteins as well.

The identified PsaR binding site is similar to operator sequences of PsaR homologs in other streptococcal species. S. gordonii ScaR binds to a similar region in the scaA promoter but also to a second inverted repeat (29). This second repeat is present in the promoter region of psaB in S. pneumoniae but not in the promoters of pcpA and prtA. SloR in S. mutans also likely exerts its repressive effect on sloABC through a larger palindromic sequence that includes the conserved region that we identify (35). Apart from the promoters of prtA, pcpA, and psaBCA, no others that contain the identified PsaR operator in their promoter regions could be found in the R6 and TIGR4 genomes. This suggests that the PsaR regulon consists of only these genes in S. pneumoniae, in contrast to the case for S. mutans, where SloR directly regulates a large number of genes (67). It is very likely that also in TIGR4 the activity of PsaR is dependent on both Zn²⁺ and Mn²⁺, since, apart from one amino acid difference (Asn161→Ser), TIGR4 PsaR is identical to the R6 and D39 PsaR sequences (data not shown). Johnston et al. (31) also found a repressive effect of Mn²⁺ and PsaR on the expression of the rlrA pathogenicity islet (SP0461 to SP0465). The rlrA locus is not present in the genomes of D39 and R6, making the effect of PsaR on rlrA and the downstream genes a strain-dependent phenomenon. There is a possible PsaR operator in the rlrA promoter region with a perfectly conserved first half-site, but a very degenerate second half-site, 5′-AAATTAAAACAACTTCCATC-3′ (consensus bases are in bold). Point mutations in the conserved bases of the first half of the operator destroyed PsaR-dependent regulation for the pcpA promoter (Table 8). However, we did not test the effect of mutations in the second half-site. Therefore, it cannot be excluded that an operator consisting of an intact first repeat and a degenerate inverted repeat, as is the case for the putative operator in the rlrA promoter, is still able to confer some weak PsaR-dependent regulation. As RlrA activates expression of the genes downstream of rlrA, namely, rrgA, rrgB, rrgC, and srtB (25), weak repression of rlrA by PsaR likely explains the upregulation of the rlrA locus in the psaR mutant (31, 68).

In conclusion, this study indicates that the relative availabilities of Zn²⁺ and Mn²⁺ in the human body could modulate the expression of several virulence genes and in this way affect the outcome of infection by S. pneumoniae.